Dietary factors, including overall caloric intake, are estimated to cause 30% of all cancers in the developed world. Post-menopausal breast cancer risk, for example, is doubled by obesity. Dietary restriction (DR, long term, low calorie diets) has established efficacy in reducing neoplastic disease in laboratory rodents. We have now identified 93 components of the sera metabolome that define a DR serotype - i.e., a metabolic serotype that reflects caloric intake. Multivariate pattern recognition analysis has identified profiles that distinguish ad libitum fed (AL) and DR rats (100% accuracy in training sets, mean >90% in test sets). Initial studies of sera proteome profiles also distinguish dietary group with 100% accuracy. The robust, systemic effects of DR suggest that a subset of the metabolites in sera that distinguish diet will reflect, even across species, the physiological benefits of reduced caloric intake. We have established an interdisciplinary, multi-site team of investigators to test the two-part hypothesis that 1) serum metabolite profiles that reflect long-term caloric intake in rats will predict relative risk of human disease (specific test case: breast cancer), and; 2) the components of these profiles are limited to a small subset of discrete metabolic pathways. Work on the latter sets the stage for understanding metabolic pathways involved in protective effects of DR against cancer. The three Aims are: Aim 1. To complete characterization of metabolic serotypes based on both the proteome and the metabolome, and to adapt these profiles for human epidemiological studies. Both individual serum constituents (proteins and metabolites) and their cognate metabolic profiles will be validated for studies using human samples. Aim 2. To determine the extent to which metabolic profiles that are reflective of or independent of long-term caloric intake predict breast cancer in nested case control studies (Hypothesis, part 1) We will evaluate the ability of expert systems trained to recognize serotypes reflecting long term caloric intake to determine relative risk for future breast cancer in sera from 1000 paired cases and controls from well-characterized human populations (Nurses' Health Study). Aim 3. To determine the biochemical identity of critical serum markers (Hypothesis, part 2) Mass spectroscopy will be used to identify critical protein and small molecule components of metabolic serotypes that characterize DR in rats and relative cancer risk in humans.